Soccer Team Performance Forecasting using artificial neural network

Abstract

Recently a large effort was spent on forecasting the outcome of sporting events. Due to forecasting perspective, the presence of competition introduces particular modeling challenges, which in turn limit the applicability of standard techniques. The objective of this study is to create a soccer team performance-forecasting model based on Artificial Neural networks that is capable of forecasting soccer players’ performance depending on teams’ history and behavior in previous matches as an input. The proposed model was trained and tested using a dataset including the features of Egypt Telecommunications club 15 years soccer team participating in the Egyptian Football Association Youth Dorian. Simulation results indicated that the proposed model could be classified as a stable predication model especially for soccer team’s status and performance, achieving high accuracy rate up to 95%.